Carbon dioxide lingers in the atmosphere far longer than most people assume, and it does not have a single, tidy expiration date. Roughly half of any CO2 pulse is absorbed by oceans and land within a few decades, but about a quarter of it remains airborne for more than a thousand years, and a stubborn fraction persists for tens of thousands of years until geological processes finally lock it away. That layered removal process is what makes CO2 unique among greenhouse gases and why its warming effect is often described as nearly irreversible on any timescale that matters to human civilization.
Why There Is No Single Number
When people ask how long CO2 lasts in the atmosphere, they usually expect an answer like “X years,” comparable to the way methane is commonly quoted at about 12 years. CO2 does not work that way. An individual molecule of CO2 cycles in and out of the air quickly, swapped between the atmosphere, ocean surface, and plants on a timescale of just a few years. But that cycling is a two-way street: for every molecule absorbed, another is released back. What actually matters for climate is how long an excess amount of CO2 stays elevated above the level nature can fully reabsorb. That “adjustment time” is vastly longer than the cycling time of any single molecule, and it plays out across multiple overlapping stages, each governed by a different part of the Earth system.
A common source of confusion comes from older analyses that estimated an atmospheric CO2 “residence time” of only four to five years. That figure reflects how quickly individual molecules swap in and out of the atmosphere, not how quickly a surplus is permanently removed.1ScienceDirect (Elsevier) / Energy. Atmospheric CO2 residence time and the carbon cycle The distinction is critical. Imagine a bathtub with the tap running and the drain open: water molecules pass through quickly, but if you increase the flow from the tap, the water level rises and stays high until the drain can catch up. For CO2, “the drain catching up” involves processes that range from decades to hundreds of millennia.
The Fast Phase: Land and Ocean Uptake
The quickest removal happens through the planet’s two major carbon sinks: land ecosystems and the surface ocean. Together, they currently absorb a little over half of what humanity emits each year. The fraction that stays in the atmosphere, known as the airborne fraction, has hovered around 44 to 47 percent over the past six decades.2Nature Communications. A regression-based approach to the CO2 airborne fraction 3Journal of Geophysical Research: Biogeosciences. Quantification of the Airborne Fraction of Atmospheric CO2 Reveals Stability in Global Carbon Sinks Over the Past Six Decades In other words, for every ton of CO2 we release, roughly half a ton is still sitting in the air a year or two later. The rest has been pulled down by trees, soils, and seawater.
Forests are the backbone of the land sink. A global synthesis of in situ forest data found that forests absorbed a steady 3.5 to 3.6 billion metric tons of carbon per year across the 1990s, 2000s, and 2010s.4PubMed. The enduring world forest carbon sink Over the decade from 2007 to 2016, the broader terrestrial sink (forests plus other vegetation and soils) removed about 3.6 billion tons of carbon annually, accounting for roughly a third of total human emissions.5Annual Review of Environment and Resources. The Terrestrial Carbon Sink The ocean handles most of the rest, absorbing CO2 at the sea surface wherever the concentration of dissolved CO2 in seawater is lower than in the air above it.6PubMed Central. Satellite estimation of global air sea CO2 flux from 2000 to 2020
This fast uptake is why, in modeling experiments that release a one-time pulse of CO2, the atmospheric concentration drops sharply during the first few decades. But the decline slows down dramatically after that initial burst of absorption, because the easy sinks fill up. Surface ocean waters become more saturated with dissolved CO2, reducing the gradient that drives further absorption. Trees and soils reach a rough balance between carbon they are gaining through growth and carbon they are losing through decomposition and disturbance. The result is a curve that drops steeply at first and then flattens into a long, slow tail.
Centuries in the Deep Ocean
Once the surface ocean has soaked up what it can, the next stage involves mixing that carbon into the deep ocean. This is not a fast process. The thermohaline circulation, the global conveyor belt of ocean currents driven by differences in temperature and salinity, flushes through the deep ocean on a timescale of roughly 900 years.7Marine Geochemistry. Thermohaline Circulation As surface water cools near the poles, it becomes denser and sinks, carrying dissolved CO2 with it into the deep. Eventually that water resurfaces elsewhere, but the full loop takes centuries. This slow conveyor is what pulls the atmospheric CO2 level down further over hundreds of years, but it cannot do the job all at once.
Additional chemistry helps. As dissolved CO2 reacts with carbonate minerals on the ocean floor and with bicarbonate ions in seawater, it is converted into forms that are effectively locked away from the atmosphere. But even this buffering process operates over centuries to millennia, not years or decades. The upshot is that a substantial fraction of excess CO2 can ride out several human lifetimes in the atmosphere before the deep ocean and its chemistry have finished their work.
The Thousand-Year Tail
Multi-model analyses of what happens to a large pulse of CO2 show a consistent and sobering pattern. After the fast sinks grab their share in the first few decades, and the deep ocean slowly works through more over centuries, roughly a quarter of the original pulse remains airborne after a full millennium. One multi-model comparison found that for a 100-billion-ton carbon pulse, about 24 percent was still in the atmosphere after 1,000 years, with the ocean having absorbed around 60 percent and land taking up the rest.8Atmospheric Chemistry and Physics. Carbon dioxide and climate impulse response functions for the computation of greenhouse gas metrics: A multi-model analysis That remaining quarter is not going anywhere fast. CO2 displays what researchers have called exceptional persistence, rendering its warming nearly irreversible for more than 1,000 years.9PubMed Central. Persistence of climate changes due to a range of greenhouse gases
This is the part that surprises most people. A gas that seems invisible and ordinary turns out to commit the planet to elevated temperatures for longer than most civilizations have existed. Even if emissions stopped overnight, the CO2 already in the air would keep warming the planet for centuries, and the oceans would continue rising for even longer as they slowly absorbed heat from the warmer atmosphere.
Geological Processes and the Final Cleanup
Removing that last stubborn fraction of excess CO2 requires geological-scale processes. Two dominate: silicate rock weathering and organic carbon burial.
Silicate weathering is essentially a very slow chemical reaction between CO2, rainwater, and certain minerals in rock (particularly those rich in calcium and magnesium). Rain dissolves atmospheric CO2 to form a weak acid, which then reacts with exposed rock, and the products are eventually washed into the ocean where the carbon is locked away in sedimentary carbonate. This process currently consumes somewhere between 150 and 330 million tons of CO2 per year, with about half of that drawdown occurring in active mountain belts where fresh rock is constantly exposed by tectonic uplift.10PubMed Central. A framework for predicting global silicate weathering and CO2 drawdown rates over geologic time-scales Compared to the roughly 37 billion tons of CO2 humans emit each year, you can see why weathering alone cannot keep up. It is the planet’s thermostat, but it operates on a timescale of tens to hundreds of thousands of years.
The other geological sink is organic carbon burial: dead plant and animal matter that sinks to the seafloor and gets trapped in sediments before it can decompose and release its carbon back. This is the largest long-term atmospheric CO2 sink after silicate weathering.11AGU Advances. Distribution and Drivers of Organic Carbon Sedimentation Along the Continental Margins Submarine landslides and turbidity currents play a surprisingly important role here, flushing large volumes of organic-rich sediment from coastal shelves into the deep ocean where it gets buried.12Geology. Globally significant mass of terrestrial organic carbon efficiently transported by canyon-flushing turbidity currents Like weathering, this process moves enormous quantities of carbon when measured over geological time, but it is far too slow to meaningfully dent a modern emissions spike on any human-relevant schedule.
What Happened the Last Time Earth Had a CO2 Spike
The closest natural analogue to what we are doing now happened about 56 million years ago during the Paleocene-Eocene Thermal Maximum, or PETM. A massive release of carbon (likely from volcanic activity, methane hydrates, or both) drove rapid warming. Sea surface temperatures spiked, ecosystems were disrupted, and extinctions accelerated. The planet eventually recovered, but it took roughly 100,000 years for the Earth system to draw down the excess CO2 and return temperatures to their pre-event levels.13Climate Change. Abrupt Climate Change: The PETM
The recovery mechanisms were exactly the slow geological ones described above. Isotopic evidence from the period shows that global erosion rates roughly doubled or tripled, and chemical weathering of rock increased by 50 to 60 percent compared to pre-warming values.14PubMed Central. Lithium isotope evidence for enhanced weathering and erosion during the Paleocene-Eocene Thermal Maximum Warmer temperatures and a more vigorous water cycle accelerated these reactions, while enhanced organic carbon burial on continental shelves contributed to the final recovery of the carbon cycle.15Earth and Planetary Science Letters. Temperature changes across the Paleocene-Eocene Thermal Maximum – a new high-resolution TEX86 temperature record from the Eastern North Sea Basin In short, nature did eventually clean up the mess, but the cleanup took longer than the entire span of modern human existence.
The PETM is an imperfect analogy because the rate of carbon release during that event was almost certainly slower than today’s emissions. The Earth’s buffering systems had more time to respond in parallel with the release. We are essentially running the same experiment at a faster pace, which may mean the atmosphere accumulates a higher peak concentration before those geological feedbacks can kick in.
Could the Sinks Weaken Over Time
One of the more worrying questions in climate science is whether the natural sinks that currently absorb half our emissions will keep doing so. The evidence so far is mixed. The airborne fraction has been remarkably stable over the past six decades, which means that sinks have been scaling up roughly in proportion to our growing emissions.3Journal of Geophysical Research: Biogeosciences. Quantification of the Airborne Fraction of Atmospheric CO2 Reveals Stability in Global Carbon Sinks Over the Past Six Decades But modeling studies suggest this cannot continue indefinitely.
Under a high-warming scenario, the terrestrial carbon sink is projected to saturate by the end of this century. One analysis found that the land sink, which absorbed about 0.96 billion tons of carbon per year in the recent baseline period, could shrink to just 0.60 billion tons per year under continued high emissions, because warming and reduced CO2 fertilization effects undercut the ability of plants and soils to keep absorbing more.16Global Biogeochemical Cycles. Saturation of Global Terrestrial Carbon Sink Under a High Warming Scenario Under a low-emissions pathway with aggressive mitigation, however, the sink actually grows. The trajectory depends almost entirely on how much we emit.
Nutrient limitations add another layer of concern. Plants need nitrogen and phosphorus to grow, and many ecosystems are already running short. Modeling that accounts for nitrogen limitation shows the land’s ability to absorb extra CO2 could be reduced by 18 to 54 percent, potentially adding 15 to 57 extra parts per million of CO2 to the atmosphere by 2100 and an additional 0.16 to 0.25°C of warming. Under high-emission scenarios, the land could even flip from a net carbon sink to a net source.17Communications Earth & Environment. Nitrogen limitation amplifies future warming by weakening terrestrial carbon cycle feedbacks and sink capacity If that happens, the share of each ton of CO2 that stays in the atmosphere would rise, and the effective atmospheric lifetime of our emissions would stretch even longer.
How CO2 Compares to Other Greenhouse Gases
Methane is often contrasted with CO2 because it is a much more potent greenhouse gas molecule for molecule, but it breaks down in the atmosphere within about a dozen years through reactions with hydroxyl radicals. Nitrous oxide lasts roughly a century before it is destroyed in the stratosphere. Both of these gases have identifiable chemical destruction pathways that remove them from the atmosphere outright. CO2 has no such pathway. It is not destroyed by sunlight or atmospheric chemistry. It can only be removed by being physically or chemically absorbed by another part of the Earth system: dissolved in the ocean, taken up by a plant, locked into rock. Every one of those removal routes is a transfer, not a destruction, and transfers can reverse if conditions change.
This is why scientists sometimes say that CO2 is “forever” in practical terms. A burst of methane causes intense short-term warming but dissipates within decades. A burst of CO2 causes less warming per molecule but commits the planet to that warming for centuries to millennia. Over very long time horizons, the cumulative climate impact of CO2 dwarfs that of shorter-lived gases, which is a key reason climate policy focuses so heavily on decarbonization.
What This Means for Carbon Removal
The multi-timescale persistence of CO2 has direct implications for the growing field of carbon dioxide removal. Any technology or natural approach that pulls CO2 out of the atmosphere but stores it in a form that could re-release the carbon within decades or centuries is fighting the same battle the ocean surface and forests are already fighting: a temporary hold, not a permanent fix. Tree planting, for example, stores carbon in biomass that can burn, decompose, or be harvested. Soil carbon can be released by changing land management or rising temperatures.
For carbon removal to truly offset fossil fuel emissions, the storage needs to be durable on a timescale comparable to the problem itself. Geological storage (injecting CO2 into deep rock formations) or mineralization (reacting CO2 with certain rocks to form stable carbonite minerals) can achieve that kind of permanence, essentially fast-forwarding the slow geological processes that would otherwise take tens of thousands of years. The challenge is doing it at a scale and cost that makes a meaningful dent in the roughly 40 billion tons of CO2 we add to the atmosphere each year.
Potential Wild Cards in the Earth System
Beyond the relatively well-understood carbon sinks, the Earth system contains reservoirs that could release large amounts of carbon under certain conditions. Methane hydrates, ice-like structures that trap methane in ocean sediments and permafrost, are one such reservoir. The “clathrate gun hypothesis” suggests that past episodes of rapid warming triggered massive hydrate destabilization, releasing methane that then amplified the warming.18Eos, Transactions American Geophysical Union. Methane Hydrates in Quaternary Climate Change: The Clathrate Gun Hypothesis While methane itself breaks down relatively quickly, the warming it causes in the interim could slow the ocean’s ability to absorb CO2, creating a feedback loop that extends the effective atmospheric lifetime of carbon emissions.
Permafrost thaw is a more widely acknowledged concern. Arctic soils hold enormous stores of organic carbon frozen for thousands of years. As high latitudes warm, microbial decomposition releases that carbon as CO2 and methane. This is not a removal process slowing down; it is a new source turning on, adding to the atmospheric burden that the existing sinks must handle. None of these feedbacks change the fundamental physics of how long CO2 persists, but they could change how much CO2 the atmosphere ends up holding and how hard the sinks have to work to bring it back down.